Effect of Statins on Intracerebral Hemorrhage

Outcome and Recurrence

Emilie FitzMaurice, AB; Lauren Wendell, MS; Ryan Snider, AB; Kristin Schwab, BA;
Rishi Chanderraj, BS; Cathrine Kinnecom, MS, RN; Kaveer Nandigam, MD; Natalia S. Rost, MD;
Anand Viswanathan, MD, PhD; Jonathan Rosand, MD, MS;
Steven M. Greenberg, MD, PhD; Eric E. Smith, MD, MPH, FRCPC

Background and Purpose—3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, or statins, have

been associated with improved outcome after ischemic stroke and subarachnoid hemorrhage but an increased risk of
incident intracerebral hemorrhage (ICH). We investigated (1) whether statin use before ICH was associated with
functional independence at 90 days, and (2) whether survivors exposed to statins after ICH had an increased risk of
recurrence.
Methods—We analyzed 629 consecutive ICH patients with 90-day outcome data enrolled in a prospective cohort study
between 1998 to 2005. Statin use was determined by patient interview at the time of ICH and supplemented by medical
record review. Independent status was defined as Glasgow Outcome Scale 4 or 5. ICH survivors were followed by
telephone interview every 6 months.
Results—Statins were used by 149/629 (24%) before ICH. There was no effect of pre-ICH statin use on the rates of
functional independence (28% versus 29%, P⫽0.84) or mortality (46% versus 45%, P⫽0.93). Medical comorbidities
and warfarin use were more common in statin users. Hematoma volumes were similar (median 28 cm3 in pre-ICH statin
users compared to 22 cm3 in nonusers, P⫽0.18). The multivariable-adjusted odds ratio for independent status in pre-ICH
statin users was 1.16 (95% CI 0.65 to 2.10, P⫽0.62). ICH survivors treated with statins after discharge did not have a
higher risk of recurrence (adjusted HR 0.82, 95% CI 0.34 to 1.99, P⫽0.66).
Conclusions—Pre-ICH statin use is not associated with improved ICH functional outcome or mortality. Post-ICH statin
use is not associated with an increased risk of ICH recurrence. (Stroke. 2008;39:2151-2154.)
Key Words: intracerebral hemorrhage 䡲 outcome 䡲 statins

3 -hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) re-

ductase inhibitors, or statins, have effects in addition to
decreasing cholesterol synthesis.1 Statin-treated animals have
better outcomes in experimental models of ischemic stroke2
and intracerebral hemorrhage (ICH).3,4 Human studies sug-
gest that statins are associated with better outcomes after
ischemic stroke5 and subarachnoid hemorrhage.6
By contrast, it is not known whether pre-ICH statin use
confers better outcomes in patients with ICH. Some of the
pleiotropic effects of statins, such as decreased platelet
aggregation7 and decreased thrombogenesis,8 could promote
hematoma expansion or increase the risk of recurrent hemor-
rhage. Patients randomized to atorvastatin in the Stroke
Prevention by Aggressive Reduction in Cholesterol Levels
(SPARCL) trial9 had an increased incidence of ICH, a finding
contrary to accumulated evidence from other randomized
trials that suggested no increased incidence.10
Therefore we investigated two hypotheses: (1) that statin
before the onset of ICH is associated with improved outcome
at 90 days, and (2) that ICH survivors taking statins are more
likely to have ICH recurrence.
Methods
Statins and ICH Outcome
We retrospectively analyzed data from an ongoing single center
prospective longitudinal cohort study of primary ICH.11 All patients
with a baseline admission CT scan, stored in digital DICOM format,
and determination of functional status at 90 days were eligible. For
patients in whom consent could not be obtained, medical record
information was stored in a database registry without patient iden-
tifiers. Clinical information, including medication use and dosage,
was abstracted from the medical record and supplemented by
interview.11 ICH volume was determined by computer-assisted
segmentation.11
There were 795 consecutive admissions with symptom onset
between January 1, 1998 and August 31, 2005. Baseline CT scan was

Received October 30, 2007; accepted November 28, 2007.

From the Hemorrhagic Stroke Research Program, Department of Neurology (E.F., L.W., R.S., K.S., R.C., C.K., K.N., N.S.R., A.V., J.R., S.M.G.,
E.E.S.) and the Center for Human Genetics Research (N.S.R., J.R.), Massachusetts General Hospital, Boston.
Correspondence to Dr Eric Smith, Associate Director of Acute Stroke Services, Mass General Hospital, Assistant Professor of Neurology, Harvard
Medical School, CPZ 300, Mass General Hospital, 55 Fruit Street, Boston, MA 02114. E-mail eesmith@partners.org
© 2008 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.107.508861

2151
2152 Stroke July 2008

not performed or was missing in 46/795 (6%), and other data were Table 1. Baseline Characteristics According to Presence of
missing in 19/795 subjects (3%), leaving 730 subjects with complete Statin Use Before ICH
baseline information.
Death within 30 days occurred in 251/730, leaving 479/730 Statin No Statin
30-day survivors. Consenting survivors or a proxy informant (to- Characteristic n⫽149, % n⫽480, % P
gether representing 273/479 of the 30-day survivors, 57%), were Age 72.4⫾9.4 71.9⫾12.7 0.65
interviewed by telephone at 90 days to determine the Glasgow
Outcome Scale (GOS) score. To reduce bias we also retrospectively Male 64 50 0.005
determined GOS at ⱖ90 days from the medical records of 105/206 Hypertension 89 75 0.0003
nonparticipating ICH survivors. The 101/206 registry subjects with- CHD 47 15 ⬍0.0001
out follow-up information had similar baseline characteristics as
subjects with follow-up information, except fewer had lobar ICH Diabetes 32 15 ⬍0.0001
location (P⫽0.02). Therefore, in sum, 90-day GOS was determined Atrial fibrillation 29 18 0.005
in a total of 629/730 potential subjects (86%).
Pre-ICH cognitive impairment 15 15 0.99

Statins and Risk of Recurrence in ICH Survivors Ischemic stroke 23 14 0.01

ICH survivors at 90 days were recruited into a prospective longitu- Previous ICH 9 6 0.19
dinal cohort study designed to find predictors of ICH recurrence.12 Warfarin 38 18 ⬍0.0001
Of 273 eligible survivors, 229 participated (82%). Subjects were
followed by telephone every 6 months.12 Dates of initiation and GCS 12 (6, 15) 12 (6, 15) 0.83
discontinuation of statins were determined based on subject inter- ICH location
view supplemented by available medical records. Dose information Deep hemispheric 44 48 0.42*
was not collected. When the date of statin initiation or discontinua-
tion could not be recalled, it was assigned as the midpoint between Lobar 44 41
successive interviews.12 Brainstem 2 4
Cerebellum 10 8
Statistical Analysis
ICH volume, cm3 28 (8, 61) 22 (8, 54) 0.15
Independent status was defined as GOS 4 or 5. Logistic regression
models were constructed to determine whether pre-ICH statin use Intraventricular extension 56 53 0.64
was associated with independent status or mortality. ICH volume 90-day independence (GOS 4 or 5) 28 29 0.84
was log-transformed, because of a nonnormal distribution, when
analyzed as the dependent variable in multivariable linear regression. 90-day mortality 46 45 0.93
Age, ICH volume, and Glasgow coma scale (GCS) score were Continuous variables displayed as mean⫾SD, or median (25th percentile,
categorized according to cut-points established in univariate analy- 75th percentile). CHD indicates coronary heart disease; ICH, intracerebral
sis, as follows: (1) age: ⱕ69, 70 to 79, ⱖ80, (2) ICH volume: 0 to hemorrhage; GCS, Glasgow Coma Scale.
29 cm3, 30 to 59 cm3, ⱖ60 cm3, (3) GCS: 3 to 10, 11 to 14, 15. *Chi-square test for distribution of hemorrhages across all locations.
In ICH survivors univariate Cox regression models and Kaplan–
Meier plots were used to determine subject characteristics associated
with an increased hazard of recurrence. Post-ICH statin use was (n⫽25), and found similar results (data not shown). To test
analyzed as a time-dependent variable because some subjects started for a dose-response effect, those with dose information
or discontinued statins during the follow-up period. Post-ICH statin (105/149, 70%) were dichotomized according to whether the
use and any variables associated with ICH recurrence (P⬍0.20) were
entered into a multivariable Cox regression model, followed by statin dose was ⱖ50% of the maximum dose recommended
backward elimination of nonsignificant variables (P⬎0.05). Statis- by the manufacturer’s labeling. Users of higher-dose statins
tical analyses were performed using SAS version 9.1.3 (SAS (n⫽25), compared to nonusers (n⫽480), were not more likely
Institute, North Carolina). to have independent status (P⫽0.78). Finally, we failed to
confirm a hypothesis that statins would have a better effect in
Results subjects with higher GCS (GCS ⬎10) or smaller ICH
Pre-ICH Statin Use and ICH Outcome volumes (⬍30 cm3), tested using interaction terms in the
Pre-ICH statin use was not associated with independent status regression models (P⬎0.50 for both).
or mortality (Table 1). Multivariable logistic regression
models showed that pre-ICH statin use was not associated Post-ICH Statin Use and Risk of Recurrence in
with independent status after adjustment for potential con- ICH Survivors
founders (Table 2). Pre-ICH statin use was associated with a Statins were used after ICH discharge in 79/229 (35%) of
nonsignificant 19% increase in ICH volume on admission CT participating ICH survivors; 57 were discharged on statins
(95% CI ⫺7% to ⫹51%, P⫽0.16) in a linear regression and 22 started statins after discharge. Mean follow-up was
model controlling for other variables associated with ICH 1.91⫾1.58 years, with a total of 437.5 person-years of
volume (male sex, hypertension, diabetes, previous ICH, and follow-up and 140.8 person-years of post-ICH statin expo-
ICH location). sure. Post-ICH statin users were more likely to be older, male,
We considered whether the effect of statins use might vary to have been on warfarin before the index ICH, and to have
according to outcome definition or patient subgroups. There CHD and diabetes (P⬍0.05 for all comparisons).
was no difference in 30-day or 90-day survival in pre-ICH Recurrent ICH occurred in 26 subjects (11%). In univariate
statin users (data not shown). Because statin withdrawal may Cox regression analysis, lobar ICH location and history of
worsen stroke outcome13 we performed additional analyses additional ICH before the index event were the only predic-
subtracting subjects in whom statins were discontinued tors of recurrence (P⬍0.05). Post-ICH statin exposure was
(n⫽22), or adding subjects in whom statins were started not associated with recurrence in univariate Cox regression
 FitzMaurice et al
Table 2. Multivariable Models of the Effect of Statin Use
Before ICH on 90-Day Independent Status (GOS 4 or 5)
OR 95% CI P Predictor
Model A 1.22 0.76 to 1.95 0.40
Adjusted for baseline characteristics
(see legend)
Model B 1.16 0.65 to 2.10 0.62
Model A⫹adjusted for ICH
characteristics (see legend)
Model C 1.24 0.62 to 2.49 0.54
Model B but restricted to
supratentorial ICH volume ⬍60
cm3, without deep coma or pre-ICH
cognitive impairment (n⫽314)
Three logistic regression models were constructed. In Model A variables in
addition to pre-ICH statin use were the pre-ICH characteristics also associated
with independent status in univariate analysis (P⬍0.20): age, coronary heart
disease, ischemic stroke, pre-ICH cognitive impairment, atrial fibrillation, and
pre-ICH warfarin use. Model B includes the same variables as Model A with the
addition of characteristics of the hemorrhagic stroke associated with indepen-
dent status in univariate analysis (P⬍0.20): GCS, ICH volume, and presence of
intraventricular hemorrhage. The 2 models were constructed separately
because hemorrhagic stroke characteristics may themselves have been
affected by pre-ICH statin use. Finally a third model was constructed with the
same variables as the model B, but excluding subjects with negative prognostic
features: deep coma (GCS 3 to 5), pre-ICH cognitive impairment, brainstem or
cerebellar location, or ICH volume ⬎60 cm3. The subject population of the
Model C was therefore similar to that studied in recent clinical trials of ICH.
GOS indicates Glasgow Outcome Scale; ICH, intracerebral hemorrhage.
(HR 0.77, 95% CI 0.32 to 1.86, P⫽0.56) or when adjusting
for other predictors (Table 3).
Discussion
We did not detect a difference in outcomes in patients taking
statins before ICH, even after adjustment for medical comor-
bidities. By contrast, treatment with statins improved senso-
rimotor recovery in 2 animal models of ICH.3,4 There may be
biological differences, however, between animals and hu-
mans in the type of injury that occurs after ICH. In contrast to
the animal studies, our subjects were taking various types and
doses of statins. A recent study suggested that statins are
associated with decreased 30-day mortality but not improved
functional outcome.14 In our larger study pre-ICH statin use
had no effect on 30-day mortality. Whether the different
results occurred because of the play of chance, or differences
in the populations studied, is not clear.
A randomized trial in persons with stroke showed that 80
mg of atorvastatin per day increased the risk of ICH.9 We did
not, however, find an increased risk of recurrence in ICH
survivors treated with statins, but note that the current study
has only modest power to detect a small increase in recur-
rence risk. In our study various statins were used, at lower
dosages than in the SPARCL trial,9 which could have led to
different results. The relationship between cholesterol levels
and ICH recurrence could not be determined in our study
because there was no systematic assessment of lipid levels.
Our results are consistent with a population-based case-
control study of ICH that concluded statin use in community
practice was not associated with an increased risk of ICH.15
Statins and Intracerebral Hemorrhage 2153
Table 3. Multivariable Cox Regression Model of ICH
Recurrence
OR 95% CI P
Post-ICH statin use 0.82 0.34 to 1.99 0.66
Lobar ICH location 2.40 1.04 to 5.57 0.04
Additional history of ICH before index event 3.26 1.12 to 9.55 0.03
Exposure to statin was entered as a time-dependent covariate to account for
statin use during only part of the follow-up period in some subjects. ICH
indicates intracerebral hemorrhage.
This main limitation of our study is that statin use was not
randomly assigned, and therefore associations between statins
and either ICH outcome or recurrent ICH may be confounded
by other factors associated with statin use, even though we
adjusted for recognized confounders using regression models.
Our analysis of the dose-response effect of statins on ICH
outcome uses the simplistic assumption that maximum dose
strengths are equipotent across different drugs.
These data come from a relatively large series of consec-
utive ICH cases with functional outcome data and prospective
follow-up and therefore provide a best estimate of the effect
of statins on ICH outcome in the absence of large trials. We
calculated that a trial of statins to improve 90-day outcome
would need to include 1026 subjects (randomized 1:1 to statin
or placebo) to have 90% power to detect an OR ⱖ1.24,
derived from Model C (Table 3), for independent status in the
statin arm. Our study suggests that patients already taking
statins for prevention of cardiovascular disease may safely
continue them during and after ICH. For ICH survivors our
data suggest that post-ICH statin exposure is not associated
with a large increase in risk of recurrence, at least with the
statin types and dosages used in community practice.
Sources of Funding
Dr Smith is supported by a grant from the National Institute of
Neurological Disorders and Stroke (K23 NS046327). Dr Rosand has
received research support from the National Institute of Neurological
Disorders and Stroke and the American Heart Association. Dr
Greenberg has received research support from the National Institute
of Neurological Disorders and Stroke and the National Institute of
Aging.
Disclosures
None.
References
1. Sacco RL, Liao JK. Drug insight: Statins and stroke. Nat Clin Pract
Cardiovasc Med. 2005;2:576 –584.
2. Endres M, Laufs U, Huang Z, Nakamura T, Huang P, Moskowitz MA,
Liao JK. Stroke protection by 3-hydroxy-3-methylglutaryl (HMG)-CoA
reductase inhibitors mediated by endothelial nitric oxide synthase. Proc
Natl Acad Sci U S A. 1998;95:8880 – 8885.
3. Jung KH, Chu K, Jeong SW, Han SY, Lee ST, Kim JY, Kim M, Roh JK.
HMG-CoA reductase inhibitor, atorvastatin, promotes sensorimotor
recovery, suppressing acute inflammatory reaction after experimental
intracerebral hemorrhage. Stroke. 2004;35:1744 –1749.
4. Seyfried D, Han Y, Lu D, Chen J, Bydon A, Chopp M. Improvement in
neurological outcome after administration of atorvastatin following
experimental intracerebral hemorrhage in rats. J Neurosurg. 2004;101:
104 –107.
5. Marti-Fabregas J, Gomis M, Arboix A, Aleu A, Pagonabarraga J, Belvis
R, Cocho D, Roquer J, Rodriguez A, Garcia MD, Molina-Porcel L,
2154 Stroke July 2008
Diaz-Manera J, Marti-Vilalta JL. Favorable outcome of ischemic stroke
in patients pretreated with statins. Stroke. 2004;35:1117–1121.
6. Tseng MY, Hutchinson PJ, Czosnyka M, Richards H, Pickard JD, Kirk-
patrick PJ. Effects of acute pravastatin treatment on intensity of rescue
therapy, length of inpatient stay, and 6-month outcome in patients after
aneurysmal subarachnoid hemorrhage. Stroke. 2007;38:1545–1550.
7. Yokoyama S, Ikeda H, Haramaki N, Yasukawa H, Katoh A, Imaizumi T.
HMG-CoA reductase inhibitor protects against in vivo arterial thrombosis
by augmenting platelet-derived nitric oxide release in rats. J Cardiovasc
Pharmacol. 2005;45:375–381.
8. Asahi M, Huang Z, Thomas S, Yoshimura S, Sumii T, Mori T, Qiu J,
Amin-Hanjani S, Huang PL, Liao JK, Lo EH, Moskowitz MA. Protective
effects of statins involving both eNOS and tPA in focal cerebral ischemia.
J Cereb Blood Flow Metab. 2005;25:722–729.
9. Amarenco P, Bogousslavsky J, Callahan A III, Goldstein LB, Hennerici
M, Rudolph AE, Sillesen H, Simunovic L, Szarek M, Welch KM, Zivin
JA. High-dose atorvastatin after stroke or transient ischemic attack.
N Engl J Med. 2006;355:549 –559.
10. Baigent C, Keech A, Kearney PM, Blackwell L, Buck G, Pollicino C,
Kirby A, Sourjina T, Peto R, Collins R, Simes R. Efficacy and safety of
cholesterol-lowering treatment: prospective meta-analysis of data from
90,056 participants in 14 randomised trials of statins. Lancet. 2005;366:
1267–1278.
11. Flibotte JJ, Hagan N, O’Donnell J, Greenberg SM, Rosand J. Warfarin,
hematoma expansion, and outcome of intracerebral hemorrhage. Neurology.
2004;63:1059–1064.
12. Viswanathan A, Rakich SM, Engel C, Snider R, Rosand J, Greenberg SM,
Smith EE. Antiplatelet use after intracerebral hemorrhage. Neurology. 2006;
66:206–209.
13. Blanco M, Nombela F, Castellanos M, Rodriguez-Yanez M, Garcia-Gil
M, Leira R, Lizasoain I, Serena J, Vivancos J, Moro MA, Davalos A,
Castillo J. Statin treatment withdrawal in ischemic stroke: a controlled
randomized study. Neurology. 2007;69:904 –910.
14. Naval NS, Abdelhak TA, Zeballos P, Urrunaga N, Mirski MA, Car-
huapoma JR. Prior statin use reduces mortality in intracerebral hemor-
rhage. Neurocrit Care. 2007.
15. Woo D, Kissela BM, Khoury JC, Sauerbeck LR, Haverbusch MA,
Szaflarski JP, Gebel JM, Pancioli AM, Jauch EC, Schneider A, Klein-
dorfer D, Broderick JP. Hypercholesterolemia, HMG-CoA reductase in-
hibitors, and risk of intracerebral hemorrhage: a case-control study.
Stroke. 2004;35:1360 –1364.